BlogBack

The finish line is just the start

Person between two Sequoia trees

Here I sit, at the halfway point between the first production Sequoia Evergreen deployment and the second — a moment over a year in the making.

We’ve been designing and building the Sequoia infrastructure since the late spring of 2013 with the goal of supporting our Equinox-hosted customers on the most fault tolerant, innovative, and consistent deployment environment possible.  It started as a bit of a skunkworks project of mine inspired by lessons we learned hosting large and complex Evergreen deployments for some very large consortia.  I saw that we were spending too much time fighting the same fires, and knew there must be a better way.  And now, there is.

The automation infrastructure — the “cloud”, if you will — behind Sequoia insists on a fairly radical departure from the traditional deployment of software like Evergreen, FulfILLment, and Koha.  The way in which it is radical, though, might seem surprising.  It is not a more complicated and tightly integrated uber-tool, not a magical monolith constructed of buzz words and hype and to-be-delivered APIs that will save us all.  No, it is, at its core, a return to the philosophy that allows Evergreen to serve the needs of both the smallest and largest of libraries.  It’s a recommitment to the UNIX programming philosophy that values  modularity, reusability, and composability.  The parts that make up the Sequoia infrastructure each do one thing, do that one thing well, and interact with one another in a predictable, composable way.  But much more important than that, Sequoia gets out of the way and allows the parts that make up Evergreen, FulfILLment, and Koha do the same; they let your ILS (and ILL) be great.

What can Sequoia do for your library, and what does it do for us?  It will take us a day or so to extract all the customizations in the exiting setup that supports the second customer that will move to Sequoia.  After that, it will take about an hour to deploy the same customer, in parallel, on our Evergreen 2.5 Sequoia cluster.  Then, once the DNS switch is flipped, they, and our support team, will no longer have to worry about the environment getting in the way of excellent uptime and customer (theirs and ours) service.

It’s true, I’m biased,  but I got my first proof that the sky’s the limit with Sequoia on its very first day hosting a production Evergreen instance when I read an email saying simply, “Seems faster, too, excellent!”

We at Equinox are all very proud of our new Sequoia Service Platform, both for its design and the promise it holds for the future, and we invite you to join us in that future.